• 検索結果がありません。

Growth property of singular solutions of linear partial differential equations in the complex domain in $C^{d+1}$ (Complex Analysis and Microlocal Analysis)

N/A
N/A
Protected

Academic year: 2021

シェア "Growth property of singular solutions of linear partial differential equations in the complex domain in $C^{d+1}$ (Complex Analysis and Microlocal Analysis)"

Copied!
4
0
0

読み込み中.... (全文を見る)

全文

(1)

大内 忠 (上智大)

Growth property of singular solutions of linear partial

differential equations in the complex domain in $C^{d+1}$

By

Sunao

$\overline{\mathrm{o}}_{\mathrm{U}\mathrm{c}}\mathrm{H}\mathrm{I}$

(Sophia Univ.)

ABSTRACT. Let $P(z, \partial)$ be a linear partial differential operator

with coefficients holomophic in a neighbourhood $\Omega$ of $z=0$ in

$C^{d+1}$. Consider equation $P(z, \partial)u(z)=f(z)$, where$u(z)$ and $f(z)$

admit singularities on the surface $\{z_{0}=0\}$. We assume that

$|f(z)|\leq A|z_{0}|^{C}$ in a region $\Omega(\theta)$ which is sectorial with respect

to$z_{0}$. The main result of this paper is the following:

There is an exponent $\gamma^{*}$ such that for some class of operators

if $\forall\epsilon>0\exists C_{\epsilon}$ such that $|u(Z)|\leq c_{\epsilon}\exp(\epsilon|Z0|^{-\gamma^{*}})$ in $\Omega(\theta)$, then $|u(Z)|\leq C|z_{0}|^{\mathrm{C}^{;}}$ for some constants $c’$ and $C$.

First

we

give the notations briefly. The coordinates of $C^{d+1}$ are

de-noted by $z=$ $(z_{0,1}z, \cdots , z_{d})=(Z_{0}, Z’)\in C\cross C^{d}$. $|z|= \max\{|Z_{i}|;0\leq$

$i\leq d\}$ and $|z’|= \max\{|Z_{i}|; 1 \leq i\leq d\}$. Its dual variables are

$\xi=(\xi_{0}, \xi’)=(\xi_{0}, \xi_{1}, \cdots, \xi_{d})$

.

$N$ is the set of all nongegative

inte-gers $N=\{0,1,2, \cdots\}$. The differentiation is denoted by $\partial_{i}=\partial/\partial z_{i}$,

and $\partial=$ $(\partial_{0}, \partial_{1}, \cdots , \partial_{d})=(\partial_{0}, \partial’)$. For a multi-indes $\alpha=(\alpha_{0}, \alpha’)\in$

$N\cross N^{d},$ $| \alpha|=\alpha_{0}+|\alpha’|=\sum_{i=0^{\alpha}}^{d}i$. Define $\partial^{\alpha}=\prod_{i=0}^{d}\partial^{\alpha}ii$. We denote $\partial^{J\alpha’}=\prod_{i=1}^{d}\partial^{\alpha_{i}}i$ by $\partial^{\alpha’}$

.

We define spaces of holomorphic functions in some regions to state

the results. Let $\Omega=\Omega_{0}\cross\Omega’$ be

a

polydisk with $\Omega_{0}=\{z_{0}\in C^{1}$; $|z_{0}|<$ $R\}$ and $\Omega’=\{z’\in C^{d};|z’|<R\}$ for

some

positive constant $R$. Put

$\Omega_{0}(\theta)=\{z_{0}\in\Omega_{0}-\{0\}; |\arg_{Z_{0}}|<\theta\}$ and $\Omega(\theta)=\Omega_{0}(\theta)\cross\Omega’$.

$\mathcal{O}(\Omega)(\mathcal{O}(\Omega’), \mathcal{O}(\Omega(\theta)))$ is the set of all holomorphic functions on

$\Omega$ (resp. $\Omega’,$ $\Omega(\theta)$) . $\mathcal{O}(\Omega(\theta))$ contains multi-valued functions, if $\theta>\pi$.

We introduce $\mathcal{O}_{(\kappa)}(\Omega(\theta))$ and $Asy_{\{\kappa\}}(\Omega(\theta))$, which

are

subspaces of

$O(\Omega(\theta))$ and fundamental function spaces in this paper.

Definition 1. $\mathcal{O}_{(\kappa)}(\Omega(\theta))(0<\kappa<+\infty)$ is the set

of

all $u(z)\in$ $\mathcal{O}(\Omega(\theta))$ such that

for

any $\epsilon>0$ and any $\theta’$ with $0<\theta’<\theta$

(1) $|u(_{Z)|}\leq C\exp(\epsilon|z_{0}|^{-}\kappa)$ $z\in\Omega(\theta/)$

holds

for

a constant $C=C(\epsilon, \theta’)$. We put $O_{(+\infty)}(\Omega(\theta))=\mathcal{O}(\Omega(\theta))$

for

$\kappa=+\infty$.

Definition 2. $\mathcal{O}_{reg,c}(\Omega(\theta))$ is the set

of

all

$u(z.)\in O(\Omega(\theta))$ such

that any $\theta’$ with $0<\theta’<\theta$

(2) $|u(_{Z)|}\leq C|_{Z}0|^{c}$ $z\in\Omega(\theta/)$

holds

for

a

constant $C=C(\theta’)$.

We say that $u(z)\in O(\Omega(\theta))$ is slowly increasing in $\Omega(\theta)$, if$u(z)\in$

$\cup$ $O_{reg,c}(\Omega(\theta))$.

$|c|<+\infty$

数理解析研究所講究録

(2)

Now let $P(z, \partial)$ be an m-th order linear partial differential equation

with coeficients in $O(\Omega)$

(3) $P(z, \partial)=\sum a\alpha(z)\partial\alpha=\sum_{m|\alpha|\leq m|\alpha|\leq}z_{0^{\alpha}}bj(\alpha z)\partial^{\alpha}$,

where $j_{\alpha}\in N$ is the valuation of $a_{\alpha}(z)$ with respect to $z_{0},$ $a_{\alpha}(z)=$

$z_{0}^{j\alpha}b_{\alpha}(z)$. Let

us

define

some

quantities for $P(z, \partial)$:

(4)

$\{$

$e_{*}:= \min\{j_{\alpha}-\alpha_{0;}\alpha\in Nd+1\},$ $\triangle=\{\alpha\in N^{d+}1j_{\alpha}-\alpha_{0=}e*\}$; $k^{*}:= \max\{|\alpha|;\alpha\in\triangle\}$.

Put

(5) $\mathfrak{P}(z, \partial)=\sum_{\alpha\in\triangle}z_{00(Z’}^{j\alpha}b,)\partial^{\alpha}\mathrm{Q}$.

Let

us

introduce

an

index which plays an important role in this paper.

Definition 3. (Minimal irregularity)

(6) $\{_{\gamma^{*}}^{\gamma^{*}}.\cdot.\cdot=\min_{=\infty},$$\{\frac{j_{\alpha}-\alpha_{0^{-e_{*}}}}{ifk=m|\alpha|-k^{*}} ; \alpha*.\in Nd+1, |\alpha|>k^{*}\}$

,

if

$k^{*}<m$,

Let us introduce conditions

on

$P(z, \partial)$.

Condition $0$

.

If

$\alpha=(\alpha_{0}, \alpha)’\in\triangle$, then $\alpha’=(0,0, \cdots , 0)$. The following condition is

more

strict than Condition $0$

.

Condition 1. $P(z, \partial)$

satisfies

Condition $\mathit{0}$ and

$b_{(k^{\mathrm{x}},0},0,\cdots,0$)(0) $\neq 0$

.

Suppose that $P(z, \partial)$ satisfies Condition $0$. Then $\mathfrak{P}(z, \partial)$ is an

ordi-nary differential operator,

(7) $\mathfrak{P}(z, \partial)=\sum_{\alpha\in\triangle}Z_{0^{*}\alpha,0(Z’}^{e}b)_{Z^{\alpha 0}}0\partial_{0}\alpha_{0}$ ,

and $\{z_{0}=0\}$ isregular singular. Define the indicial polynomial$\chi_{P}(z’, \lambda)$

$\mathrm{o}\mathrm{f}\mathfrak{P}(Z, \partial)$,

(8) $\chi_{P}(_{Z’}, \lambda):=\sum b_{\alpha,0}(z)\lambda(\lambda-1)\cdots(\lambda-\alpha_{0}+1)\alpha\in\triangle/$ .

Further suppose that $P(z, \partial)$ satisfies Condition 1. Then $\chi_{P}(Z’, \lambda)$ is

a

polynomial of $\lambda$ with degree $k^{*}$ in $\{z;|z|\leq R\}$. Hence there exist

real constants $a_{0},$$a_{1}$ and $b_{0}$ such that all the roots of$\chi_{P}(z’,$ $\lambda \mathrm{I}=0$ for

$|z|\leq R$

are

contained in $\{\lambda;a_{0}\leq\Re\lambda\leq a_{1}, |^{\alpha}- s\lambda|\leq b_{0}\}$.

Now let

us

consider

(Eq) $P(_{Z}, \partial)u(z)=f(z)$.

We have results concerning the growth properties of solutions of (Eq).

(3)

Theorem 4. Suppose that$P(z, \partial)$

satisfies

Condition 1. Let $u(z)\in$

$\mathcal{O}_{(\gamma^{*})(\Omega())}\theta$ be

a

solution

of

(Eq). Suppose that $f(z)\in O_{reg,c}(\Omega(\theta))$.

Thenthere is

a

polydisk$U$ centered at$z=0$ such that $u(z)\in \mathcal{O}_{re_{\mathit{9}^{c’}}},(U(\theta))$

for

any $c’< \min\{c-e_{*}, a_{0}\}$.

Theorem 5. Suppose that $P(z, \partial)$

satisfies

Condition $\mathit{0}$

.

Let $u(z)\in$

$O_{(\gamma^{*})}(\Omega(\theta))$ be a solution

of

(Eq). Suppose that $f(z)\in \mathcal{O}_{reg,c}(\Omega(\theta))$.

Then there is

a

polydisk $U$ centered at $z=0$ and a constant $c$” such

that $u(z)\in \mathcal{O}_{reg,\mathrm{c}^{l\prime}}(U(\theta))$.

We show Theorem 4 by constructing a parametrix and Theorem 5

follows from Theorem 4. The proof theorems and the details of this

paper will be appeared in the forthcoming paper.

We give

some

examples satisfying Condition 1:

(a). Operators of normal type with respect to $\partial_{0}$,

$\partial_{0}^{k^{*}}+\alpha 0<k\sum*a(\alpha z)\partial^{\alpha}$.

(b). Operators of Fuchsian type.

(c). Other concrete examples

are

$I_{d}+z_{00}^{2}\partial+z_{0}\partial^{2}1$

’ $z0\partial^{2}0^{+(_{Z)\partial_{0}}}a+\partial 13$. The present paper follows $\overline{\mathrm{O}}$

uchi [4]. The class of operators

consid-ered in [4] was more strict than that of this paper. The main Theorem

in [4] was the following:

If

$u(z)$ grows at most some exponential order near $z_{0}=0_{\mathrm{Z}}$ that $is_{f}$

for

any $\epsilon>0$ $|u(z)|\leq C_{\epsilon}\exp(\epsilon|z0|-\gamma^{*})$ near $z_{0}=0$, and

if

$f(z)$

hehaves asymptotically $f(z) \sim\sum_{n=0}^{+}\infty f_{n}(Z’)z_{0}n$ as $z_{0}arrow \mathit{0}$ in a sectorial

region $\Omega(\theta)$, where $|f_{n}(z’)|\leq AB^{n}\Gamma(n/\gamma^{*}+1)$, then $u(z)$ has also the

asymptotic expansion like $f(z)a\mathit{8}Z_{0}$ tends to $\mathit{0}$.

It

was

an extension of the main result of [1] and [2]. But in the

present paper we treat a wider class of operators which contains that

of [4]. So even if $f(z)$ has

a

Gevrey type asymptotic expansion, $u(z)$

does not always have. Hence, Theorem 4 in this paper is somewhat

different. Roughly speaking,

if

$u(z)$ grows at most some exponential order near $\{z_{0}=0\}$, and

if

$f(z)$ has the $sl_{\mathit{0}}wl,y$ increasing singularities on $\{z_{0}=0\}_{f}$ then the

growth order

of

singularities

of

$u(z)$ are also slowly increasing.

We

can

show the results in [4], by using Theorem 4.

REFERENCES [1] $\overline{\mathrm{O}}$

uchi,S., An integral representation ofsingular solutions and removable

sin-gularities tolinearpartial

differential

equations, Publ. RIMS KyotoUniv. 26,

735-783 (1990).

(4)

[2] $\overline{\mathrm{O}}$uchi, S., The behaviour

of

solutions with singularities on a characteristic

surface

to linear partial

differential

equations in the complex domains, Publ.

RIMS Kyoto Univ. 29, 63-120 (1993).

[3] $\overline{\mathrm{O}}$

uchi,S., Genuine solution8 and

formal

solution8 with Gevrey type estimates

ofnonlinearpartialdefferntial equations, J. Math. Sci. Univ. Tokyo 2 (1995),

375-417.

[4] $\overline{\mathrm{O}}$uchi,

S., Singular solutions with asymptotic $ex.panSi_{\mathit{0}}n$

of

linear partial

dif-ferential equations in the complex domain

Sunao $\overline{\mathrm{o}}_{\mathrm{U}\mathrm{c}}\mathrm{H}\mathrm{I}$

Department of Mathematics, Sophia University

Kioicho Chiyoda-ku, Tokyo 102,

JAPAN

参照

関連したドキュメント

Because of the restriction of differential equations, we obtain that the properties of fixed points of meromorphic solutions of higher order linear differential equations

ABSTRACT: The decomposition method is applied to examples of hyperbolic, parabolic, and elliptic partlal differential equations without use of linearlzatlon techniques.. We

This is the continuation of an investigation of basic boundary value problems for first order complex model partial differential equations.. Model second order equations are the

In the present paper, it is shown by an example that a unit disc counterpart of such finite set does not contain all possible T- and M-orders of solutions, with respect to

Solvability conditions for linear differential equations are usually formulated in terms of orthogonality of the right-hand side to solutions of the homogeneous adjoint

We recall here the de®nition of some basic elements of the (punctured) mapping class group, the Dehn twists, the semitwists and the braid twists, which play an important.. role in

This paper is concerned with the existence, the uniqueness, convergence and divergence of formal power series solutions of singular first order quasi-linear partial

BELAïDI, Estimation of the hyper-order of entire solutions of complex linear ordinary differential equations whose coefficients are entire func- tions, E. Qualitative Theory